 Okay, welcome back to SiliconANGLE Wikibon's theCUBE, our flagship program. We've got a lot of the events, they strike the signal from the noise. We are here at our exclusive OpenStack Enterprise Forum, put on by the great folks here in the cloud business, OpenStack business, we're excited. All the vendors are here, SolidFire, HP, big event. SolidFire put this event on theCUBE is here. I'm John Furrier with Dave Vellante, co-founder of wikibon.org and SiliconANGLE's theCUBE. Our next guest is Rodney Peck, who's an OpenStack cloud architect. Oh, JC Martin, JC Martin. He said Rodney. I said, I'm sorry. JC Martin, cloud architect, eBay. We just had Rodney on, sorry about that. JC, welcome to theCUBE. Thanks. So, what is the big deal going on with eBay and OpenStack? Take us through your implementation and what you guys have done. Sure. So, our journey to cloud started a while back and we developed our own OpenStack management software. But like two years ago, we looked at the landscape and we tried to see if there was a way to leverage an open source technology. We looked at various vendors or projects like CloudStack and Amazon, Azure. And what we really found is that the community that was starting to grow on OpenStack was definitely something that was appealing. So, what we did is we evaluated OpenStack and started deploying it in two specific use cases. One was developer productivity. So, giving developers virtual machines instead of using their desktops. And this proved to be very popular. So, today, for example, we have 7,000 VMs created by developers. More than 5,000 of them are using this infrastructure to innovate, build new products without having to request new infrastructure or do some kind of ticketing, use some ticketing system to get machines or resource. We looked also at another project which was how do we accelerate the time to site, time to release small projects that can be changing user experience, some kind of experimentation project. So, we developed like an internal public cloud where business units or developers could experiment and deploy just a small application, try it out, see how it works. Sandbox. Sandbox isolated from the rest of the site some experience that they had on public cloud basically providing the same experience that they would get on Amazon, but on our infrastructure. So, giving like the public cloud experience internally. And that has proven to be very popular. And at the same time on the PayPal side, they took OpenStack and they deployed in production OpenStack, trying to accelerate time to deliver applications to site. And that also has been very popular. So, same time frame, two years ago, we took some two different paths, but right now we are looking at consolidating those two tracks, if you want, into one common cloud and reap out the benefits of aligning and sharing infrastructure. So, essentially, you're duplicating the benefits of the public cloud. You talked about the Amazon public cloud, but you're doing it in-house. I think it was Chris Kemp said that every Amazon customer that he talks to knows that crossover point when bringing it in-house is going to be less expensive than doing it in the public cloud. I wanted to ask you as a practitioner, is that true, did you have that sense? Were you guys such scale that you already felt as though bringing it in-house was more advantageous economically or were there other factors involved? Were there other capabilities that you've been able to develop that you maybe couldn't get from the Amazon cloud? What was the decision point there? So, we did the evaluation on how much it would cost us to move to a public cloud. And as Chris was saying, there's an inflection point where if you go above, it's going to be much more expensive than what you would get internally. So, it all depends on how spiky your demand is, right? So, for example, at eBay, we have seasonal traffic around the end of the year. And if we look at what is our baseline traffic and we can move part of that excess capacity to a public cloud, then obviously, it's more interesting than moving everything because you have this baseline that you always need that maybe you want to keep internally and not pay rent on it. But for things that you need only a few times of the year, it makes sense economically to have- So, do you burst to the public cloud? Not yet, but we are evaluating options to burst because this has an architectural impact, right? Applications have to be ready to burst to the cloud. So, there's the infrastructure. So, we are building the infrastructure. We are experimenting always with cloud provider to see how this would work. But we have to talk to our developers and say what happens if you are on a public cloud, not in our data center. So, you're talking about the user experience, recovery, backup, all the- Security, latency. Compliance. Exactly. So, we have to look at all those aspects. The technical side is easy. That's not a big deal. Is it all those? The technical side is easy. You know, you connect the cloud to another cloud. That's easy. JC, I want to ask you about the enterprise. Obviously, tonight, it's a packed house. People online are going crazy. A lot of popularity open stack. So, Dave and I were commenting earlier yesterday, the hyperscale mindsets bleeding over to the enterprise fast. So, you're seeing the mindset of DevOps aggressively hit the mainstream enterprise. Yeah. Why is open stack such a good fit there, in your opinion? Given someone who's playing with it and deploying it? Because it's open source. So, you can easily look under the hood, right? So, if you are buying a commercial product, you are in the same type of model that you used to be with legacy infrastructure and big vendors. With open stack, the new breed of DevOps administrators, they have the opportunity to even contribute to the software, right? They have a community that they can talk to when they have an issue. It's a career pass also, because if you put open stack in your resume, you will see that you will get several job offers per week. And you put Python, open stack, DevOps, and you are golden, right? So, it's a career pass, and the community is the key reason why I think... So, you're seeing people who have open stack skills are getting jobs pretty fast. And we cannot hire them anymore because they are such a big demand. They are such a big demand. That's awesome. Dave, what's your take on that? What's your view on DevOps, hires, in mainstream enterprise? I mean, DevOps, they eat glass and spit nails. Well, but I think that the discussion, actually I thought Lydia Leong said it up well, there's two vectors in private cloud. One is extending virtualization, it's automation, and the other is the DevOps piece. And I think my take on it is, if I'm a young developer, I want to go into the DevOps piece. So, that's really the action is, what kind of, what databases are you using for your open stack applications? One of each. Why, all of the above? Yes, for open stack itself or on open stack? Yeah, so, well, let's talk about the databases that you're using for applications running on open stack. Okay, so, we just recently released a database as a service using the Trove project from open stack. So, we have several developers that are contributors to that project. And it's using MySQL primarily because that's what was the main, if you want, database supported by this project. But we are actively looking at supporting no SQL options because we have a large community of developers that want to have MongoDB and others. So, for us, it's about giving choice to developers so that they can innovate on top of the platform. And what we have seen is that we may have started with Oracle, everybody was going to Oracle database because it was the easiest choice for them because we had support from our platform directly on Oracle. But as we offer more choice, and it's easy for them to deploy their own instances, we are seeing a shift where people are now trying and developing new architectures that are supporting those new databases. So, you're providing services much in the same way that AWS is providing services. We are trying to. You got a relational database, you got plans for no SQL, you got Oracle. In fact, our developers, as someone else was saying, they were raised with the public cloud. They know services that are in AWS and they are expecting the same things from us. Last question. What's your next OpenStack project? Next OpenStack, we have many of those. We are trying to push for better load balancer as a service, firewall as a service. We are contributing into Solom, platform as a service, pushed by Rackspace, and we hope that this is going to be one of the OpenStack projects, moving up the stack from infrastructure and service to platform. Awesome, JC, thanks very much. Appreciate it. Welcome. Thanks for coming on, JC. We really appreciate it. We love the commentary. OpenStack is the future for the enterprise. You can look under the hood. That by itself is the table stakes for enterprises. And by the way, customization, Lego blocks, however you want to look at it, whatever your description, open source, looking under the hood. At the end of the day, something has to be supported and that's what people want. This is theCUBE, we'll be right back after this short break.